A digital intelligent management method and system for line loss in a transformer area

Abstract

This invention proposes a digital and intelligent management method and system for transformer substation line loss. The method includes: real-time acquisition of power parameters at each measuring point in the transformer substation; construction of a support vector machine model, and detection of line loss anomalies and anomaly patterns based on real-time power parameter data; grey analysis combined with the detection results of anomaly patterns to evaluate the energy efficiency status value of power parameters at each measuring point, and comprehensive analysis of the energy efficiency status value of individual parameters using fuzzy set theory to obtain the overall line loss energy efficiency index for the entire transformer substation; prediction of the line loss energy efficiency index for a future preset time period; analysis of the predicted line loss energy efficiency index using multiple decision theories to formulate multiple loss reduction decisions, and multi-criteria grey decision analysis to obtain the optimal loss reduction decision scheme. This invention, through digital and intelligent management of transformer substation line loss, not only improves the efficiency and accuracy of line loss management but also effectively reduces line loss risk and optimizes transformer substation operation strategies.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a digital intelligent management method and system for line losses in a transformer substation. Background Art

[0002] In the power system, the line loss in the substation refers to the power loss caused by various reasons (such as line resistance, equipment loss, power theft, etc.) in the substation (i.e. the power supply range of the distribution transformer) during the transmission of power from the power station to the user. Line loss not only affects the economic benefits of the power system, but also may pose a threat to the safe and stable operation of the power grid. Therefore, the management and optimization of substation line loss is an important part of the operation of the power system.

[0003] At present, the line loss management in the substation mainly relies on traditional measurement and statistical methods, such as regular meter reading, manual inspection, etc. These methods have the following defects: due to the errors of manual meter reading and inspection, as well as the lag of data processing, it is difficult for traditional methods to accurately assess the actual situation of line loss in the substation; traditional methods mainly rely on manual experience and simple statistical analysis, lack intelligent decision-making support means, and it is difficult to formulate scientific and reasonable loss reduction measures; traditional methods often only focus on a single line loss indicator, lack the ability to comprehensively evaluate different factors, resulting in insufficient pertinence and effectiveness of loss reduction measures. Summary of the invention

[0004] Based on this, the purpose of the present invention is to propose a digital intelligent management method and system for line loss in a substation area to solve the above-mentioned problems.

[0005] According to a digital intelligent management method for line loss in a transformer area proposed by the present invention, the method comprises:

[0006] Real-time acquisition of power parameters at each measuring point in the substation, including voltage, current, temperature and power factor;

[0007] Construct a support vector machine model and detect abnormal line loss and abnormal line loss patterns based on real-time power parameter data of the substation area;

[0008] Combined with the line loss abnormal pattern detection results, grey analysis is performed to evaluate the energy efficiency status value of the power parameters at each measuring point, and the energy efficiency status value of a single parameter is synthesized using fuzzy set theory to obtain the overall line loss energy efficiency comprehensive index of the substation area;

[0009] Based on the current and historical comprehensive line loss energy efficiency indicators of the entire substation, predict the comprehensive line loss energy efficiency indicators for the preset time period in the future;

[0010] A variety of decision-making theories are used to analyze the predicted comprehensive indicators of line loss energy efficiency in the substation to make a variety of loss reduction decisions, and multi-criteria grey decision analysis is performed on various loss reduction decisions to obtain the optimal loss reduction decision plan.

[0011] Furthermore, the step of combining the line loss abnormal mode detection result with gray analysis to evaluate the energy efficiency status value of the power parameters at each measuring point includes:

[0012] The measured values ​​of the power parameters at each measuring point are taken as an event set, denoted as X = {x1, x2, ..., x i , ..., x n}, where x i represents the measured value of the power parameter at the i-th measuring point, where n is the number of measuring points;

[0013] The level representing the energy efficiency status is taken as the target strategy set, denoted as Y = {y1, y2, ..., y j , ..., y m}, where y j represents the jth energy efficiency status level;

[0014] For each energy efficiency status level y j , define a whitening function f j (x i ), used to map the elements in the event set X to a real number in the interval [0, 1], indicating the degree of membership of the energy efficiency status level corresponding to the corresponding event;

[0015] According to the whitening function, the membership degree of each power parameter of the measuring point corresponding to each energy efficiency status level is calculated to form a grey comprehensive evaluation matrix R, where R ij Indicates the membership degree of the power parameter of the i-th measuring point corresponding to the j-th energy efficiency status level;

[0016] Based on the difference in the impact of power parameters at different measuring points on the energy efficiency of the substation, a weight is assigned to each measuring point to form a weight vector W = {w1, w2, ..., w n};

[0017] Combining the weight vector and the grey comprehensive evaluation matrix, the energy efficiency state value of the single-parameter power parameter is calculated by weighted summation. The calculation formula is:

[0018]

[0019] Among them, E i represents the energy efficiency state value of the power parameter at the i-th measuring point, R ij is the membership degree of the power parameter of the i-th measuring point corresponding to the j-th energy efficiency status level, y j is the numerical representation of the jth energy efficiency status level, w i is the weight of the i-th measurement point.

[0020] Furthermore, the step of synthesizing the energy efficiency state value of a single parameter by using fuzzy set theory to obtain the overall line loss energy efficiency comprehensive index of the substation includes:

[0021] An energy efficiency impact value EI is defined for each abnormal line loss mode, where the value range of the energy efficiency impact value EI is from 0 to 1. The larger the EI value, the greater the impact of the corresponding abnormal line loss mode on the energy efficiency of the substation area.

[0022] The single parameter energy efficiency state values ​​of all measurement points are regarded as a fuzzy set;

[0023] For each measuring point, the single parameter energy efficiency state value Ei is weighted according to the energy efficiency impact value EI corresponding to its related line loss abnormal mode. The formula is:

[0024] E i ′=E i (1-EI k ),

[0025] Among them, E i ′ is the weighted energy efficiency status value, EI k is the energy efficiency impact value corresponding to the abnormal line loss mode associated with the i-th measuring point;

[0026] According to the weighted energy efficiency state value E i ′ and weight vector W, calculate the overall line loss energy efficiency comprehensive index EC of the substation area, the formula is:

[0027] Furthermore, the steps of constructing a support vector machine model and detecting line loss anomalies and line loss anomaly patterns based on real-time power parameter data of the substation area include:

[0028] Clean the historical power parameter data of the substation area, process missing values ​​and abnormal values, and extract features related to line loss anomalies;

[0029] The extracted features related to line loss anomalies are used as input, and the line loss anomalies and their abnormal pattern labels are used as output to build a support vector machine model. During the construction, the support vector machine model is trained with the training set data, and the optimal decision boundary is found by solving the optimization problem to maximize the interval between samples of different line loss abnormal patterns.

[0030] The constructed support vector machine model receives the power parameter data of the substation in real time to detect the real-time line loss anomaly and its abnormal mode, and outputs the detection results.

[0031] Furthermore, the step of predicting the comprehensive line loss energy efficiency index for a preset time period in the future according to the current and historical comprehensive line loss energy efficiency index of the entire substation includes:

[0032] Construct a time series prediction model and train it with historical data sets so that the model can learn the law of changes of line loss energy efficiency comprehensive indicators over time;

[0033] Construct time series data based on the comprehensive indicators of line loss energy efficiency of the current and historical substations;

[0034] The time series data is input into a trained time series prediction model to predict the comprehensive index of line loss energy efficiency in a future preset time period.

[0035] Furthermore, the step of inputting the time series data into a trained time series prediction model to predict the comprehensive index of line loss energy efficiency in a future preset time period includes:

[0036] The LSTM layer of the model captures the long-term dependencies in the time series data. Based on the prediction results of the previous time step, the hidden state of each time step in the future time period is updated and output. The output of the LSTM layer is: t+1 =LSTM(a t ,h t ), where a t is the input data of the current time step, h t is the hidden state of the current time step, h t+1 is the hidden state at the next time step;

[0037] At each time step, features are extracted from the output of the LSTM layer through the fully connected layer of the model to generate a prediction for the next time step: z t+1 =Dense(h t+1 ), where z t+1 is the predicted value for the next time step;

[0038] The predicted values ​​of each time step are combined to obtain the comprehensive index of line loss energy efficiency in the predicted future time period.

[0039] Furthermore, the step of performing multi-criteria grey decision analysis on various loss reduction decisions to obtain the optimal loss reduction decision solution includes:

[0040] The loss reduction decision schemes derived from various decision theories are used as the strategy set;

[0041] Establish an event set based on various influencing factors of the comprehensive index of line loss energy efficiency in the substation area, including grid structure, equipment aging, load characteristics and environmental factors;

[0042] Determine decision-making objectives, including reducing line loss rates in substations and improving energy efficiency;

[0043] Obtain the effect value of each countermeasure under each event to construct the effect sample matrix of each decision-making target;

[0044] By effect measure transformation, the effect sample matrix is ​​converted into a consistent effect measure matrix;

[0045] Allocate decision-making power according to the importance of each decision-making goal;

[0046] The consistent effect measurement matrix and decision weight of each decision target are weighted and summed to obtain the comprehensive effect measurement matrix;

[0047] According to the comprehensive effect measurement matrix, the loss reduction decision plan with the best comprehensive effect is selected as the final plan.

[0048] The present invention also proposes a digital intelligent management system for line loss in a transformer area, the system comprising:

[0049] Data acquisition module: used to obtain the power parameters of each measuring point in the substation in real time, including voltage, current, temperature and power factor;

[0050] Anomaly detection module: used to build a support vector machine model and detect line loss anomalies and line loss abnormality patterns based on real-time power parameter data of the substation area;

[0051] Line loss energy efficiency evaluation module: It is used to combine the line loss abnormal mode detection results to perform gray analysis to evaluate the energy efficiency status value of the power parameters at each measuring point, and use fuzzy set theory to synthesize the energy efficiency status value of a single parameter to obtain the overall line loss energy efficiency comprehensive index of the substation area;

[0052] Prediction module: used to predict the comprehensive line loss energy efficiency index in the future preset time period based on the current and historical comprehensive line loss energy efficiency index of the entire substation area;

[0053] Decision-making module: It is used to analyze the predicted comprehensive indicators of line loss energy efficiency of the substation using a variety of decision-making theories to formulate a variety of loss reduction decisions, and to conduct multi-criteria grey decision analysis on various loss reduction decisions to obtain the optimal loss reduction decision plan.

[0054] In summary, according to the above-mentioned digital intelligent management method of line loss in a substation, the power parameters of each measuring point in the substation are obtained in real time, and by building a support vector machine model, line loss anomalies are detected and abnormal patterns are identified to quickly locate the source of the problem, and then gray analysis is performed in combination with the line loss abnormal pattern detection results to evaluate the energy efficiency status value of the power parameters at each measuring point. This evaluation method can comprehensively consider multiple factors, such as the incompleteness and uncertainty of data, so as to obtain more accurate and reliable energy efficiency evaluation results, and use fuzzy set theory to synthesize the energy efficiency status values ​​of single parameters to obtain the overall line loss energy efficiency comprehensive index of the substation. This process handles fuzzy and uncertain information, making the comprehensive evaluation results more in line with the actual situation. Then, based on the current and historical comprehensive line loss energy efficiency indicators of the entire substation, the comprehensive line loss energy efficiency indicators of the future preset time period are predicted, which is convenient for formulating loss reduction measures in advance and optimizing the substation operation strategy to reduce future line loss risks. Then, a variety of decision-making theories are used to analyze the predicted comprehensive line loss energy efficiency indicators of the substation to formulate a variety of loss reduction decisions, and multi-criteria gray decision analysis is performed on various loss reduction decisions to comprehensively consider multiple factors, so as to obtain the optimal loss reduction decision plan. The present invention not only improves the efficiency and accuracy of substation line loss management through digital management of substation line loss, but also effectively reduces line loss risks and optimizes substation operation strategies.

[0055] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description or will be learned through embodiments of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0057] Figure 1 This is a flow chart of a digital intelligent management method for line loss in a transformer area according to the first embodiment of the present invention;

[0058] Figure 2 This is a system block diagram of a digital intelligent management system for line losses in a substation according to the second embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0060] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0061] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0062] Embodiment 1

[0063] See also Figure 1 The present invention proposes a digital intelligent management method for line loss in a transformer area, the method comprising steps S101 to S105:

[0064] S101, real-time acquisition of power parameters of each measuring point in the substation, including voltage, current, temperature and power factor.

[0065] Smart meters can collect power parameters at various monitoring points (measurement points) in the substation in real time. These parameters include but are not limited to voltage, current, temperature and power factor, etc. Among them, voltage and current are the basic data for evaluating the operating status of the power system, temperature reflects the working environment of the equipment and possible overheating problems, and power factor is related to the utilization efficiency of electric energy.

[0066] S102, constructing a support vector machine model, and detecting line loss anomalies and line loss anomaly patterns based on real-time power parameter data of the substation area.

[0067] Using the support vector machine model, based on the real-time collected power parameter data, the line loss anomalies and their abnormal patterns in the substation area are detected. Support vector machines are good at classification and regression analysis, and are particularly suitable for processing nonlinear problems and high-dimensional data. Through the support vector machine model, power parameter combinations that do not conform to the normal state can be identified and classified, so as to determine whether there are line loss anomalies and identify the corresponding line loss abnormal patterns, in preparation for subsequent energy efficiency evaluation and loss reduction decisions.

[0068] S103, gray analysis is performed in combination with the line loss abnormal pattern detection results to evaluate the energy efficiency status value of the power parameters at each measuring point, and the energy efficiency status value of the single parameter is integrated using fuzzy set theory to obtain the overall line loss energy efficiency comprehensive index of the substation.

[0069] Combined with the detection results of the abnormal line loss mode, gray analysis is performed to evaluate the energy efficiency status value of the power parameters at each measuring point. When gray analysis is used here to evaluate the energy efficiency status value of the power parameters at each measuring point, it can calculate the energy efficiency status value of each measuring point by considering the relationship between the power parameter measurement value and the energy efficiency status level, as well as the difference in the impact of different measuring points on the energy efficiency of the substation. Through gray analysis, the present invention can perform energy efficiency evaluation on the power parameters of each measuring point, thereby providing a direction for improving energy efficiency.

[0070] Then, the energy efficiency status values ​​of these single parameters are combined using fuzzy set theory to obtain the overall line loss energy efficiency comprehensive index of the substation. First, the energy efficiency status values ​​of the power parameters of each measuring point can be fuzzified and converted into fuzzy sets. Fuzzy sets allow parameter values ​​to fluctuate within a certain range, which can better reflect the uncertainty of the actual situation. Then, considering the influence of multiple parameters, the energy efficiency status values ​​of the power parameters of each measuring point can be comprehensively evaluated through the fuzzy comprehensive evaluation method, and the overall line loss energy efficiency comprehensive index of the substation can be obtained through fuzzy operation.

[0071] S104, predicting the comprehensive line loss energy efficiency index for a preset time period in the future based on the current and historical comprehensive line loss energy efficiency index for the entire substation area.

[0072] Historical data can be used to establish a time series model to predict future comprehensive indicators of line loss energy efficiency. Different prediction time periods can be set according to actual needs, such as day, week, month, etc. This embodiment facilitates the formulation of loss reduction measures in advance through accurate prediction of future comprehensive indicators of line loss energy efficiency, thereby optimizing resource allocation.

[0073] S105, using a variety of decision-making theories to analyze the predicted comprehensive indicators of line loss energy efficiency in the substation to formulate a variety of loss reduction decisions, and performing multi-criteria grey decision analysis on various loss reduction decisions to obtain the optimal loss reduction decision plan.

[0074] In implementation, a variety of decision-making theories (such as cost-benefit analysis, risk analysis, multi-objective optimization, etc.) can be used to analyze the predicted comprehensive indicators of line loss energy efficiency in the substation area, so as to comprehensively consider all aspects of loss reduction decisions and formulate a variety of loss reduction decisions. And through multi-criteria grey decision analysis, the optimal loss reduction decision plan is selected. When choosing among multiple loss reduction decision plans, multiple criteria (such as cost, benefit, risk, etc.) are comprehensively considered through grey decision analysis to deal with these complex decision problems. Through comprehensive analysis, the optimal decision plan that can effectively reduce line losses and meet other constraints (such as cost, feasibility, etc.) is selected.

[0075] Based on steps S101 to S105, the power parameters of each measuring point in the substation are obtained in real time, and the support vector machine model is constructed to detect line loss anomalies and identify abnormal patterns to quickly locate the source of the problem. Gray analysis is then performed in combination with the line loss abnormal pattern detection results to evaluate the energy efficiency status values ​​of the power parameters at each measuring point. This evaluation method can comprehensively consider multiple factors, such as the incompleteness and uncertainty of the data, so as to obtain more accurate and reliable energy efficiency evaluation results, and the energy efficiency status values ​​of the single parameters are synthesized using fuzzy set theory to obtain the overall line loss energy efficiency comprehensive index of the substation. This process handles fuzzy and uncertain information, so that the comprehensive evaluation results are more in line with the actual situation. Then, based on the current and historical overall line loss energy efficiency comprehensive indicators of the substation, the line loss energy efficiency comprehensive indicators of the future preset time period are predicted, so as to facilitate the formulation of loss reduction measures in advance and optimize the substation operation strategy to reduce future line loss risks. Then, a variety of decision theories are used to analyze the predicted line loss energy efficiency comprehensive indicators of the substation to formulate a variety of loss reduction decisions, and a multi-criteria gray decision analysis is performed on various loss reduction decisions to comprehensively consider multiple factors, so as to obtain the optimal loss reduction decision plan. The present invention not only improves the efficiency and accuracy of line loss management in the substation through digital management of line loss in the substation, but also effectively reduces the risk of line loss and optimizes the operation strategy of the substation.

[0076] The following is a further detailed introduction to a digital intelligent management method for line loss in a transformer area according to an embodiment of the present invention:

[0077] Further optionally, in step S103, the step of performing grey analysis on the line loss abnormal mode detection result to evaluate the energy efficiency status value of the power parameter at each measuring point includes:

[0078] The measured values ​​of the power parameters at each measuring point are taken as an event set, denoted as X = {x1, x2, ..., x i , ..., x n}, where x i represents the measured value of the power parameter at the i-th measuring point, where n is the number of measuring points;

[0079] The level representing the energy efficiency status (such as excellent, good, average, poor) is taken as the target countermeasure set, denoted as Y = {y1, y2, ..., y j , ..., y m}, where y j Indicates the j-th energy efficiency status level (e.g., excellent = 1, good = 0.75, average = 0.5, poor = 0.25);

[0080] For each energy efficiency status level y j , define a whitening function f j (x i), used to map the elements in the event set X to a real number in the interval [0, 1], indicating the degree of membership of the energy efficiency status level corresponding to the corresponding event;

[0081] According to the whitening function, the membership degree of each power parameter of the measuring point corresponding to each energy efficiency status level is calculated to form a grey comprehensive evaluation matrix R, where R ij Indicates the membership degree of the power parameter of the i-th measuring point corresponding to the j-th energy efficiency status level;

[0082] Based on the difference in the impact of power parameters at different measuring points on the energy efficiency of the substation, a weight is assigned to each measuring point to form a weight vector W = {w1, w2, ..., w n};

[0083] Combining the weight vector and the grey comprehensive evaluation matrix, the energy efficiency state value of the single-parameter power parameter is calculated by weighted summation. The calculation formula is:

[0084]

[0085] Among them, E i represents the energy efficiency state value of the power parameter at the i-th measuring point, R ij is the membership degree of the power parameter of the i-th measuring point corresponding to the j-th energy efficiency status level, y j is the numerical representation of the jth energy efficiency status level, w i is the weight of the i-th measurement point.

[0086] Understandably, the energy efficiency status value of the power parameters of each measuring point is evaluated by combining the detection results of the abnormal line loss mode with gray analysis. First, the measured values ​​of the power parameters of each measuring point constitute the event set X, and the level of the energy efficiency status constitutes the target countermeasure set Y. Then, a whitening function is defined for each energy efficiency status level to map the elements in the event set to real numbers in the interval [0, 1], indicating the membership of the corresponding energy efficiency status level. The membership of the power parameters of each measuring point corresponding to each energy efficiency status level is then calculated by the whitening function to form a gray comprehensive evaluation matrix R. Then, based on the difference in the impact of the power parameters of different measuring points on the energy efficiency of the substation, weights are assigned to each measuring point to form a weight vector W. Finally, the energy efficiency status value of the power parameters of each measuring point is calculated by weighted summation in combination with the weight vector and the gray comprehensive evaluation matrix, which takes into account the membership, the numerical representation of the energy efficiency status level, and the weight of each measuring point. The single-parameter energy efficiency status value of this embodiment is calculated through grey analysis. By referring to the membership and weights of multiple energy efficiency status levels, the fuzziness or uncertainty of the energy efficiency status can be well handled, avoiding the one-sidedness of single indicator evaluation, thereby more comprehensively reflecting the energy efficiency status of the power parameters of the measuring point, improving the accuracy of the evaluation, and by accurately evaluating the energy efficiency status values ​​of the power parameters of each measuring point, the problem of low energy efficiency can be discovered and handled in a timely manner, thereby reducing the loss of the power system.

[0087] Further optionally, in step S103, the step of synthesizing the energy efficiency state value of a single parameter using fuzzy set theory to obtain the overall line loss energy efficiency comprehensive index of the substation includes:

[0088] An energy efficiency impact value EI is defined for each abnormal line loss mode, where the value range of the energy efficiency impact value EI is from 0 to 1. The larger the EI value, the greater the impact of the corresponding abnormal line loss mode on the energy efficiency of the substation area.

[0089] The single parameter energy efficiency state values ​​of all measurement points are regarded as a fuzzy set;

[0090] For each measuring point, the single parameter energy efficiency state value Ei is weighted according to the energy efficiency impact value EI corresponding to its related line loss abnormal mode. The formula is:

[0091] E i ′=E i (1-EI k ),

[0092] Among them, E i ′ is the weighted energy efficiency status value, EI k is the energy efficiency impact value corresponding to the abnormal line loss mode associated with the i-th measuring point;

[0093] According to the weighted energy efficiency state value E i′ and weight vector W, calculate the overall line loss energy efficiency comprehensive index EC of the substation area, the formula is:

[0094] It can be understood that in the line loss management of the substation, the single-parameter energy efficiency status value is integrated through fuzzy set theory to obtain the overall energy efficiency index. Specifically, an energy efficiency impact value EI is defined for each line loss abnormal mode, and the value range is 0 to 1. The larger the EI value, the greater the impact of the corresponding line loss abnormal mode on the energy efficiency of the substation, which is used to quantify the impact of the line loss abnormal mode on the energy efficiency; then the single-parameter energy efficiency status value of all measuring points is used as a fuzzy set, and the fuzzy set allows the parameter value to fluctuate within a certain range (because the single-parameter energy efficiency status value of this embodiment is a real number calculated by gray analysis, this real number contains a certain fuzziness or uncertainty of the energy efficiency state). For each measuring point, the single-parameter energy efficiency status value EI is calculated based on the energy efficiency impact value EI corresponding to its related line loss abnormal mode. i Here, the influence of abnormal line loss mode on energy efficiency status value is considered, and the original energy efficiency status value is adjusted. Finally, according to the weighted energy efficiency status value E i ′ and weight vector W, calculate the overall line loss energy efficiency comprehensive index EC of the substation.

[0095] Further optionally, in step S102, the step of constructing a support vector machine model and detecting line loss anomalies and line loss anomaly patterns based on real-time area power parameter data includes:

[0096] Clean the historical power parameter data of the substation area, process missing values ​​and abnormal values, and extract features related to line loss anomalies;

[0097] The extracted features related to line loss anomalies are used as input, and the line loss anomalies and their abnormal pattern labels are used as output to build a support vector machine model. During the construction, the support vector machine model is trained with the training set data, and the optimal decision boundary is found by solving the optimization problem to maximize the interval between samples of different line loss abnormal patterns.

[0098] The constructed support vector machine model receives the power parameter data of the substation in real time to detect the real-time line loss anomaly and its abnormal mode, and outputs the detection results.

[0099] It is understandable that by constructing a support vector machine model, the deep mining and utilization of the power parameter data of the substation area is realized. First, the historical power parameter data of the substation area is cleaned, missing values ​​and abnormal values ​​are processed, and features closely related to line loss anomalies are extracted; then, these features are used as input, and the line loss anomaly and its abnormal mode are used as labels. The support vector machine model is trained through the training set data, and the optimal decision boundary is found through the optimization algorithm to maximize the interval between samples of different line loss abnormal modes, thereby improving the classification accuracy of the model; finally, the constructed model is applied to the reception and processing of real-time power parameter data of the substation area, realizing real-time detection of line loss anomalies and their abnormal modes, and outputting accurate detection results, thereby effectively improving the accuracy and real-time performance of the line loss anomaly detection results.

[0100] Further optionally, in step S104, the step of predicting the comprehensive line loss energy efficiency index for a preset time period in the future based on the current and historical comprehensive line loss energy efficiency index of the entire substation includes:

[0101] Construct a time series prediction model and train it with historical data sets so that the model can learn the law of changes of line loss energy efficiency comprehensive indicators over time;

[0102] Construct time series data based on the comprehensive indicators of line loss energy efficiency of the current and historical substations;

[0103] The time series data is input into a trained time series prediction model to predict the comprehensive index of line loss energy efficiency in a future preset time period.

[0104] Understandably, a time series prediction model can be constructed using deep learning frameworks such as TensorFlow or PyTorch, and trained through historical data sets, so that the model can capture and learn the inherent laws of changes in the comprehensive line loss energy efficiency indicators over time; then, based on the current and historical comprehensive line loss energy efficiency indicators of the substation, a complete time series data is constructed; this time series data is input into the trained prediction model to achieve accurate prediction of the comprehensive line loss energy efficiency indicators within a preset time period in the future, thereby providing an accurate basis for the formulation of subsequent loss reduction decisions, and can effectively improve the foresight and initiative of substation line loss energy efficiency management.

[0105] Further optionally, the step of inputting the time series data into a trained time series prediction model to predict the comprehensive index of line loss energy efficiency in a future preset time period includes:

[0106] The LSTM layer of the model captures the long-term dependencies in the time series data. Based on the prediction results of the previous time step, the hidden state of each time step in the future time period is updated and output. The output of the LSTM layer is: t+1=LSTM(a t ,h t ), where a t is the input data of the current time step, h t is the hidden state of the current time step, h t+1 is the hidden state at the next time step;

[0107] At each time step, features are extracted from the output of the LSTM layer through the fully connected layer of the model to generate a prediction for the next time step: z t+1 =Dense(h t+1 ), where z t+1 is the predicted value for the next time step;

[0108] The predicted values ​​of each time step are combined to obtain the comprehensive index of line loss energy efficiency in the predicted future time period.

[0109] It can be understood that the LSTM layer is used to capture the long-term dependencies in the time series data. Based on the prediction results of the previous time step, the hidden state of each future time step is dynamically updated and output, ensuring that the model can accurately capture the dynamic changes of the time series; at each time step, the key features are extracted from the output of the LSTM layer through the fully connected layer to generate the prediction value of the next time step. This process further improves the accuracy and reliability of the prediction; finally, the prediction values ​​of all time steps are integrated to form a complete prediction result, that is, the comprehensive line loss energy efficiency index for the future time period, thereby significantly improving the predictability of the line loss energy efficiency management of the substation area, and providing strong support for the stable operation and energy efficiency optimization of the power system.

[0110] Further optionally, in step S105, the step of performing multi-criteria grey decision analysis on various loss reduction decisions to obtain the optimal loss reduction decision solution includes:

[0111] The loss reduction decision schemes derived from various decision theories are used as the strategy set;

[0112] Establish an event set based on various influencing factors of the comprehensive index of line loss energy efficiency in the substation area, including grid structure, equipment aging degree, load characteristics, environmental factors, etc.;

[0113] Determine decision-making objectives, including reducing line loss rate in substations and improving energy efficiency;

[0114] Obtain the effect value of each countermeasure under each event to construct the effect sample matrix of each decision-making target;

[0115] By effect measure transformation, the effect sample matrix is ​​converted into a consistent effect measure matrix;

[0116] Allocate decision-making power according to the importance of each decision-making goal;

[0117] The consistent effect measurement matrix and decision weight of each decision target are weighted and summed to obtain the comprehensive effect measurement matrix;

[0118] According to the comprehensive effect measurement matrix, the loss reduction decision plan with the best comprehensive effect is selected as the final plan.

[0119] Understandably, multi-criteria grey decision analysis is used to scientifically and systematically determine the optimal loss reduction decision. First, the loss reduction strategies derived from different decision theories constitute a set of countermeasures in the grey group decision, and an event set is established based on various influencing factors of the comprehensive indicators of line loss and energy efficiency in the substation, and the decision goals are determined at the same time, such as reducing the line loss rate in the substation and improving energy efficiency. By collecting the effect data of each countermeasure under different events, an effect sample matrix is ​​constructed, and the grey group decision theory is used to perform effect measurement transformation to ensure the consistency of the evaluation. According to the importance of each decision goal, the decision weights are reasonably allocated, and the consistent effect measurement matrix is ​​weighted and summed to obtain a comprehensive effect measurement matrix reflecting the comprehensive effects of each countermeasure, and the loss reduction decision scheme with the best comprehensive effect is selected according to the comprehensive effect measurement matrix. This embodiment can comprehensively and objectively evaluate various loss reduction strategies through multi-criteria grey decision analysis, and screen out the best loss reduction decision, providing a scientific basis for substation line loss management and energy efficiency improvement, and ensuring the scientificity and effectiveness of the decision.

[0120] Embodiment 2

[0121] See also Figure 2 The present invention proposes a digital intelligent management system for line loss in a transformer area, the system comprising:

[0122] Data acquisition module: used to obtain the power parameters of each measuring point in the substation in real time, including voltage, current, temperature and power factor;

[0123] Anomaly detection module: used to build a support vector machine model and detect line loss anomalies and line loss abnormality patterns based on real-time power parameter data of the substation area;

[0124] Line loss energy efficiency evaluation module: It is used to combine the line loss abnormal mode detection results to perform gray analysis to evaluate the energy efficiency status value of the power parameters at each measuring point, and use fuzzy set theory to synthesize the energy efficiency status value of a single parameter to obtain the overall line loss energy efficiency comprehensive index of the substation area;

[0125] Prediction module: used to predict the comprehensive line loss energy efficiency index in the future preset time period based on the current and historical comprehensive line loss energy efficiency index of the entire substation area;

[0126] Decision-making module: It is used to analyze the predicted comprehensive indicators of line loss energy efficiency of the substation using a variety of decision-making theories to formulate a variety of loss reduction decisions, and to conduct multi-criteria grey decision analysis on various loss reduction decisions to obtain the optimal loss reduction decision plan.

[0127] Further optionally, the line loss energy efficiency evaluation module is also used for:

[0128] The measured values ​​of the power parameters at each measuring point are taken as an event set, denoted as X = {x1, x2, ..., x i , ..., x n}, where x i represents the measured value of the power parameter at the i-th measuring point, where n is the number of measuring points;

[0129] The level representing the energy efficiency status is taken as the target strategy set, denoted as Y = {y1, y2, ..., y j , ..., y m}, where y j represents the jth energy efficiency status level;

[0130] For each energy efficiency status level y j , define a whitening function f j (x i ), used to map the elements in the event set X to a real number in the interval [0, 1], indicating the degree of membership of the energy efficiency status level corresponding to the corresponding event;

[0131] According to the whitening function, the membership degree of each power parameter of the measuring point corresponding to each energy efficiency status level is calculated to form a grey comprehensive evaluation matrix R, where R ij Indicates the membership degree of the power parameter of the i-th measuring point corresponding to the j-th energy efficiency status level;

[0132] Based on the difference in the impact of power parameters at different measuring points on the energy efficiency of the substation, a weight is assigned to each measuring point to form a weight vector W = {w1, w2, ..., w n};

[0133] Combining the weight vector and the grey comprehensive evaluation matrix, the energy efficiency state value of the single-parameter power parameter is calculated by weighted summation. The calculation formula is:

[0134]

[0135] Among them, E i represents the energy efficiency state value of the power parameter at the i-th measuring point, R ij is the membership degree of the power parameter of the i-th measuring point corresponding to the j-th energy efficiency status level, y j is the numerical representation of the jth energy efficiency status level, w i is the weight of the i-th measurement point.

[0136] Further optionally, the line loss energy efficiency evaluation module is also used for:

[0137] An energy efficiency impact value EI is defined for each abnormal line loss mode, where the value range of the energy efficiency impact value EI is from 0 to 1. The larger the EI value, the greater the impact of the corresponding abnormal line loss mode on the energy efficiency of the substation area.

[0138] The single parameter energy efficiency state values ​​of all measurement points are regarded as a fuzzy set;

[0139] For each measuring point, the single parameter energy efficiency state value Ei is weighted according to the energy efficiency impact value EI corresponding to its related line loss abnormal mode. The formula is:

[0140] E i ′=E i (1-EI k ),

[0141] Among them, E i ′ is the weighted energy efficiency status value, EI k is the energy efficiency impact value corresponding to the abnormal line loss mode associated with the i-th measuring point;

[0142] According to the weighted energy efficiency state value E i ′ and weight vector W, calculate the overall line loss energy efficiency comprehensive index EC of the substation area, the formula is:

[0143] Further optionally, the above anomaly detection module is also used for:

[0144] Clean the historical power parameter data of the substation area, process missing values ​​and abnormal values, and extract features related to line loss anomalies;

[0145] The extracted features related to line loss anomalies are used as input, and the line loss anomalies and their abnormal pattern labels are used as output to build a support vector machine model. During the construction, the support vector machine model is trained with the training set data, and the optimal decision boundary is found by solving the optimization problem to maximize the interval between samples of different line loss abnormal patterns.

[0146] The constructed support vector machine model receives the power parameter data of the substation in real time to detect the real-time line loss anomaly and its abnormal mode, and outputs the detection results.

[0147] Further optionally, the above prediction module is also used for:

[0148] Construct a time series prediction model and train it with historical data sets so that the model can learn the law of changes of line loss energy efficiency comprehensive indicators over time;

[0149] Construct time series data based on the comprehensive indicators of line loss energy efficiency of the current and historical substations;

[0150] The time series data is input into a trained time series prediction model to predict the comprehensive index of line loss energy efficiency in a future preset time period.

[0151] Further optionally, the above prediction module is also used for:

[0152] The LSTM layer of the model captures the long-term dependencies in the time series data. Based on the prediction results of the previous time step, the hidden state of each time step in the future time period is updated and output. The output of the LSTM layer is: t+1 =LSTM(a t ,h t ), where a t is the input data of the current time step, h t is the hidden state of the current time step, h t+1 is the hidden state at the next time step;

[0153] At each time step, features are extracted from the output of the LSTM layer through the fully connected layer of the model to generate a prediction for the next time step: z t+1 =Dense(h t+1 ), where z t+1 is the predicted value for the next time step;

[0154] The predicted values ​​of each time step are combined to obtain the comprehensive index of line loss energy efficiency in the predicted future time period.

[0155] Optionally, the decision module is further used to:

[0156] The loss reduction decision schemes derived from various decision theories are used as the strategy set;

[0157] Establish an event set based on various influencing factors of the comprehensive index of line loss energy efficiency in the substation area, including grid structure, equipment aging, load characteristics and environmental factors;

[0158] Determine decision-making objectives, including reducing line loss rates in substations and improving energy efficiency;

[0159] Obtain the effect value of each countermeasure under each event to construct the effect sample matrix of each decision-making target;

[0160] By effect measure transformation, the effect sample matrix is ​​converted into a consistent effect measure matrix;

[0161] Allocate decision-making power according to the importance of each decision goal;

[0162] The consistent effect measurement matrix and decision weight of each decision target are weighted and summed to obtain the comprehensive effect measurement matrix;

[0163] According to the comprehensive effect measurement matrix, the loss reduction decision plan with the best comprehensive effect is selected as the final plan.

[0164] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A digital intelligent management method for line loss in a substation area, characterized in that: The method comprises: Real-time acquisition of power parameters at each measuring point in the substation, including voltage, current, temperature and power factor; Construct a support vector machine model and detect abnormal line loss and abnormal line loss patterns based on real-time power parameter data of the substation area; Combined with the line loss abnormal pattern detection results, grey analysis is performed to evaluate the energy efficiency status value of the power parameters at each measuring point, and the energy efficiency status value of a single parameter is synthesized using fuzzy set theory to obtain the overall line loss energy efficiency comprehensive index of the substation area; Based on the current and historical comprehensive line loss energy efficiency indicators of the entire substation, predict the comprehensive line loss energy efficiency indicators for the preset time period in the future; Use a variety of decision-making theories to analyze the predicted comprehensive indicators of line loss energy efficiency in the substation to formulate a variety of loss reduction decisions, and conduct multi-criteria grey decision analysis on various loss reduction decisions to obtain the optimal loss reduction decision plan; The step of combining the line loss abnormal mode detection result with gray analysis to evaluate the energy efficiency status value of the power parameters at each measuring point includes: The measured values ​​of the power parameters at each measuring point are taken as an event set, denoted as X={x1, x2, ..., x i , ..., x n }, where x i represents the measured value of the power parameter at the i-th measuring point, where n is the number of measuring points; The level representing the energy efficiency status is taken as the target strategy set, denoted as Y = {y1, y2, ..., y j , ..., y m }, where y j represents the jth energy efficiency status level; For each energy efficiency status level y j , define a whitening function f j (x i ), used to map the elements in the event set X to a real number in the interval [0, 1], indicating the degree of membership of the energy efficiency status level corresponding to the corresponding event; According to the whitening function, the membership degree of each power parameter of the measuring point corresponding to each energy efficiency status level is calculated to form a grey comprehensive evaluation matrix R, where R ij Indicates the membership degree of the power parameter of the i-th measuring point corresponding to the j-th energy efficiency status level; Based on the difference in the impact of power parameters at different measuring points on the energy efficiency of the substation, a weight is assigned to each measuring point to form a weight vector W = {w1, w2, ..., w n }; Combining the weight vector and the grey comprehensive evaluation matrix, the energy efficiency state value of the single-parameter power parameter is calculated by weighted summation. The calculation formula is: ， Among them, E i represents the energy efficiency state value of the power parameter at the i-th measuring point, R ij is the membership degree of the power parameter of the ith measuring point corresponding to the jth energy efficiency status level, y j is the numerical representation of the jth energy efficiency status level, w i is the weight of the ith measurement point.

2. The digital intelligent management method for line loss in a substation area according to claim 1 is characterized in that: The step of synthesizing the energy efficiency state value of a single parameter by using fuzzy set theory to obtain the overall line loss energy efficiency comprehensive index of the substation includes: An energy efficiency impact value EI is defined for each abnormal line loss mode, where the value range of the energy efficiency impact value EI is from 0 to 1. The larger the EI value, the greater the impact of the corresponding abnormal line loss mode on the energy efficiency of the substation area. The single parameter energy efficiency state values ​​of all measurement points are regarded as a fuzzy set; For each measuring point, the single parameter energy efficiency state value Ei is weighted according to the energy efficiency impact value EI corresponding to its related line loss abnormal mode. The formula is: AND i ′=AND i (1−EI k ), Among them, E i ′ is the weighted energy efficiency status value, EI k is the energy efficiency impact value corresponding to the abnormal line loss mode associated with the i-th measuring point; According to the weighted energy efficiency state value E i ′ and weight vector W, calculate the overall line loss energy efficiency comprehensive index EC of the substation area, the formula is: .

3. The digital intelligent management method for line loss in a substation area according to claim 1 is characterized in that: The steps of constructing a support vector machine model and detecting line loss anomalies and line loss anomaly patterns based on real-time power parameter data of the substation area include: Clean the historical power parameter data of the substation area, process missing values ​​and abnormal values, and extract features related to line loss anomalies; The extracted features related to line loss anomalies are used as input, and the line loss anomalies and their abnormal pattern labels are used as output to build a support vector machine model. During the construction, the support vector machine model is trained with the training set data, and the optimal decision boundary is found by solving the optimization problem to maximize the interval between samples of different line loss abnormal patterns. The constructed support vector machine model receives the power parameter data of the substation in real time to detect the real-time line loss anomaly and its abnormal mode, and outputs the detection results.

4. The digital intelligent management method for line loss in a substation area according to claim 1, characterized in that: The step of predicting the comprehensive line loss energy efficiency index for a preset time period in the future according to the current and historical comprehensive line loss energy efficiency index of the entire substation includes: Construct a time series prediction model and train it with historical data sets so that the model can learn the law of changes of line loss energy efficiency comprehensive indicators over time; Construct time series data based on the comprehensive indicators of line loss energy efficiency of the current and historical substations; The time series data is input into a trained time series prediction model to predict the comprehensive index of line loss energy efficiency in a future preset time period.

5. The digital intelligent management method for line loss in a substation area according to claim 4 is characterized in that: The step of inputting the time series data into a trained time series prediction model to predict the comprehensive index of line loss energy efficiency in a future preset time period includes: The LSTM layer of the model captures the long-term dependencies in the time series data. Based on the prediction results of the previous time step, the hidden state of each time step in the future time period is updated and output. The output of the LSTM layer is: t+1 =LSTM(a t ,h t ), where a t is the input data of the current time step, h t is the hidden state of the current time step, h t+1 is the hidden state at the next time step; At each time step, features are extracted from the output of the LSTM layer through the fully connected layer of the model to generate a prediction for the next time step: z t+1 =Dense(h t+1 ), where z t+1 is the predicted value for the next time step; The predicted values ​​of each time step are combined to obtain the comprehensive index of line loss energy efficiency in the predicted future time period.

6. The digital intelligent management method for line loss in a substation area according to claim 1, characterized in that: The step of performing multi-criteria grey decision analysis on various loss reduction decisions to obtain the optimal loss reduction decision solution includes: The loss reduction decision schemes derived from various decision theories are used as the strategy set; Establish an event set based on various influencing factors of the comprehensive index of line loss energy efficiency in the substation area, including grid structure, equipment aging, load characteristics and environmental factors; Determine decision-making objectives, including reducing line loss rates in substations and improving energy efficiency; Obtain the effect value of each countermeasure under each event to construct the effect sample matrix of each decision-making target; By effect measure transformation, the effect sample matrix is ​​converted into a consistent effect measure matrix; Allocate decision-making power according to the importance of each decision-making goal; The consistent effect measurement matrix and decision weight of each decision target are weighted and summed to obtain the comprehensive effect measurement matrix; According to the comprehensive effect measurement matrix, the loss reduction decision plan with the best comprehensive effect is selected as the final plan.

7. A digital intelligent management system for line loss in a substation area, characterized in that: The system comprises: Data acquisition module: used to obtain the power parameters of each measuring point in the substation in real time, including voltage, current, temperature and power factor; Anomaly detection module: used to build a support vector machine model and detect line loss anomalies and line loss abnormality patterns based on real-time power parameter data of the substation area; Line loss energy efficiency evaluation module: It is used to combine the line loss abnormal mode detection results to perform gray analysis to evaluate the energy efficiency status value of the power parameters at each measuring point, and use fuzzy set theory to synthesize the energy efficiency status value of a single parameter to obtain the overall line loss energy efficiency comprehensive index of the substation area; Prediction module: used to predict the comprehensive line loss energy efficiency index in the future preset time period based on the current and historical comprehensive line loss energy efficiency index of the entire substation area; Decision-making module: It is used to analyze the comprehensive indicators of line loss energy efficiency of the predicted substation using various decision-making theories to formulate various loss reduction decisions, and to perform multi-criteria grey decision analysis on various loss reduction decisions to obtain the optimal loss reduction decision plan; Wherein, the line loss energy efficiency evaluation module is also used for: The measured values ​​of the power parameters at each measuring point are taken as an event set, denoted as X={x1, x2, ..., x i , ..., x n }, where x i represents the measured value of the power parameter at the i-th measuring point, where n is the number of measuring points; The level representing the energy efficiency status is taken as the target strategy set, denoted as Y = {y1, y2, ..., y j , ..., y m }, where y j represents the jth energy efficiency status level; For each energy efficiency status level y j , define a whitening function f j (x i ), used to map the elements in the event set X to a real number in the interval [0, 1], indicating the degree of membership of the energy efficiency status level corresponding to the corresponding event; According to the whitening function, the membership degree of each power parameter of the measuring point corresponding to each energy efficiency status level is calculated to form a grey comprehensive evaluation matrix R, where R ij Indicates the membership degree of the power parameter of the i-th measuring point corresponding to the j-th energy efficiency status level; Based on the difference in the impact of power parameters at different measuring points on the energy efficiency of the substation, a weight is assigned to each measuring point to form a weight vector W = {w1, w2, ..., w n }; Combining the weight vector and the grey comprehensive evaluation matrix, the energy efficiency state value of the single-parameter power parameter is calculated by weighted summation. The calculation formula is: ， Among them, E i represents the energy efficiency state value of the power parameter at the i-th measuring point, R ij is the membership degree of the power parameter of the ith measuring point corresponding to the jth energy efficiency status level, y j is the numerical representation of the jth energy efficiency status level, w i is the weight of the ith measurement point.

Images